{"@context":"http://iiif.io/api/presentation/3/context.json","id":"https://ualberta.aviaryplatform.com/iiif/1j9765c059/manifest","type":"Manifest","label":{"en":["Computers and Games 2022: Keynote 3 - Tao Qin"]},"logo":"https://d9jk7wjtjpu5g.cloudfront.net/organizations/logo_images/000/000/128/original/UA_Logo_WHT_RGB_%281%29.png?1725471982","metadata":[{"label":{"en":["Description"]},"value":{"en":["\u003cp\u003eDeep Reinforcement Learning for Game Playing and Testing\u003c/p\u003e\r\n\u003cp\u003ePowered by recent advances of deep learning and reinforcement learning, game AI has made remarkable progress in recent years. In this talk, I will present two of our recent projects on deep RL for game playing and testing. (1) Project Suphx: the World-Best Mahjong AI, which demonstrated stronger performance than most top human players in terms of stable rank and is rated above 99.99% of all the officially ranked human players in the Tenhou platform. This is the first time that a computer program achieves 10 DAN and outperforms most top human players in Mahjong. (2) Project Mariana: Pixel based AI for Automated Game Testing.  We built a general game testing agent, Inspector, that is only based on screenshots/pixels and can be easily applied to different games without deep integration with games.\u003c/p\u003e\r\n\u003cp\u003eDr. Tao Qin is a Senior Principal Researcher at Microsoft Research AI4Science. His research interests include deep learning (with applications to scientific discovery, machine translation, speech synthesis and recognition), reinforcement learning (with applications to games and real-world problems), and game theory and multi-agent systems (with applications to cloud computing, online and mobile advertising). His team won the first place for 8 translation tasks in WMT 2019, built the world-best Mahjong AI, named Suphx, which achieved 10 DAN on the Tenhou platform in 2019, and developed the FastSpeech series models supporting 100+ languages and 300+ voices in Azure TTS.\u003c/p\u003e"]}},{"label":{"en":["Series or Event Name"]},"value":{"en":["Computers and Games 2022"]}},{"label":{"en":["Rights Statement"]},"value":{"en":["\u003cp\u003e\u003ca href=\"https://creativecommons.org/licenses/by/4.0/\"\u003eAttribution 4.0 International (CC by 4.0)\u003c/a\u003e\u003c/p\u003e"]}},{"label":{"en":["Date"]},"value":{"en":["2022-11-23"]}},{"label":{"en":["Type"]},"value":{"en":["Educational","Event"]}},{"label":{"en":["Agent"]},"value":{"en":["Kishimoto, Akihiro (Speaker)","Qin, Tao (Presenter)"]}},{"label":{"en":["Subject"]},"value":{"en":["Artificial Intelligence (Topical)","Computer Gaming (Topical)"]}},{"label":{"en":["Language"]},"value":{"en":["English"]}}],"summary":{"en":["\u003cp\u003eDeep Reinforcement Learning for Game Playing and Testing\u003c/p\u003e\r\n\u003cp\u003ePowered by recent advances of deep learning and reinforcement learning, game AI has made remarkable progress in recent years. In this talk, I will present two of our recent projects on deep RL for game playing and testing. (1) Project Suphx: the World-Best Mahjong AI, which demonstrated stronger performance than most top human players in terms of stable rank and is rated above 99.99% of all the officially ranked human players in the Tenhou platform. This is the first time that a computer program achieves 10 DAN and outperforms most top human players in Mahjong. (2) Project Mariana: Pixel based AI for Automated Game Testing. \u0026nbsp;We built a general game testing agent, Inspector, that is only based on screenshots/pixels and can be easily applied to different games without deep integration with games.\u003c/p\u003e\r\n\u003cp\u003eDr. Tao Qin is a Senior Principal Researcher at Microsoft Research AI4Science. His research interests include deep learning (with applications to scientific discovery, machine translation, speech synthesis and recognition), reinforcement learning (with applications to games and real-world problems), and game theory and multi-agent systems (with applications to cloud computing, online and mobile advertising). His team won the first place for 8 translation tasks in WMT 2019, built the world-best Mahjong AI, named Suphx, which achieved 10 DAN on the Tenhou platform in 2019, and developed the FastSpeech series models supporting 100+ languages and 300+ voices in Azure TTS.\u003c/p\u003e"]},"requiredStatement":{"label":{"en":["Attribution"]},"value":{"en":["\u003cp\u003e\u003ca href=\"https://creativecommons.org/licenses/by/4.0/\"\u003eAttribution 4.0 International (CC by 4.0)\u003c/a\u003e\u003c/p\u003e"]}},"provider":[{"id":"https://ualberta.aviaryplatform.com/aboutus","type":"Agent","label":{"en":["University of Alberta Library"]},"homepage":[{"id":"https://ualberta.aviaryplatform.com/","type":"Text","label":{"en":["University of Alberta Library"]},"format":"text/html"}],"logo":[{"id":"https://d9jk7wjtjpu5g.cloudfront.net/organizations/logo_images/000/000/128/original/UA_Logo_WHT_RGB_%281%29.png?1725471982","type":"Image"}]}],"thumbnail":[{"id":"https://d9jk7wjtjpu5g.cloudfront.net/collection_resource_files/thumbnails/000/248/529/small/6.Keynote3-TaoQin.mp4_1724174116.jpg?1724174119","type":"Image","format":"image/jpeg"}],"items":[{"id":"https://ualberta.aviaryplatform.com/collections/1771/collection_resources/133569/file/248529","type":"Canvas","label":{"en":["Media File 1 of 1 - 6._Keynote_3_-_Tao_Qin.mp4"]},"duration":3914.091,"width":640,"height":360,"thumbnail":[{"id":"https://d9jk7wjtjpu5g.cloudfront.net/collection_resource_files/thumbnails/000/248/529/small/6.Keynote3-TaoQin.mp4_1724174116.jpg?1724174119","type":"Image","format":"image/jpeg"}],"items":[{"id":"https://ualberta.aviaryplatform.com/collections/1771/collection_resources/133569/file/248529/content/1","type":"AnnotationPage","items":[{"id":"https://ualberta.aviaryplatform.com/collections/1771/collection_resources/133569/file/248529/content/1/annotation/1","type":"Annotation","motivation":"painting","body":{"id":"https://aviary-p-ualberta.s3.wasabisys.com/collection_resource_files/resource_files/000/248/529/original/6._Keynote_3_-_Tao_Qin.mp4?1724174116","type":"Video","format":"video/mp4","duration":3914.091,"width":640,"height":360},"target":"https://ualberta.aviaryplatform.com/collections/1771/collection_resources/133569/file/248529","metadata":[]}]}],"annotations":[]}]}